The systemic host inflammatory response to microbial invasion ("sepsis") can damage the infected host, leading to multiple organ failure, coagulopathy, shock, and death. Despite much effort, recent attempts to improve the outcome of severe sepsis in humans have been unsuccessful. In particular, using clinical criteria to select the best candidates for anti-cytokine therapy has proven very difficult: it is not possible to determine whether a patient will benefit most from none, a little, or a lot of a given antidote. The research proposed here takes a new approach to this problem. The goal of this approach is to limit (damp) the inflammatory response without interfering with the normal, beneficial defense against invading microbes. DNA constructs will be engineered so that sepsis antidote proteins (anti-endotoxins or anti-cytokines) will be expressed downstream of promoters that respond to inflammatory cytokines (e.g., IL- 1, IL-6, TNF-alpha). These constructs, introduced into mice using adenoviral vectors or DNA-liposome complexes, will be tested for their ability to protect the animals from gram-negative bacterial infection or endotoxin challenge. Since production of the antidote proteins in vivo will be controlled by the intensity of the animal's own inflammatory response, dangerous inflammatory excess may be prevented without blocking the beneficial components of the host response. In effect, each animal will determine its own dose of antidote protein(s). Potential clinical uses include the prophylaxis and therapy of sepsis and other systemic inflammatory disorders, as well as certain tissue-specific applications.